Design and Validation of Segmented CFRP Lamella-Based Composite End Shield for Bearing Current Mitigation

Abstract

This study addresses the premature failure of electric motor bearings caused by inverter-induced parasitic currents. We propose a novel segmented end shield design utilizing 24 carbon fiber-reinforced polymer (CFRP) lamellae to provide both structural support and galvanic isolation. The “main working” of the design relies on a segmented architecture where the lamellae are adhesively bonded between a central bearing housing and an outer mounting flange, creating a high-impedance path that interrupts circulating currents. Experimental validation focused on both mechanical stability and dielectric performance. Results indicate that the assembly maintains rotor positional integrity under nominal loads while providing an insulation resistance > 1 GΩ at 1 kV and a structural capacitance of 2.47 nF. These parameters effectively mitigate low-frequency circulating currents. Data analysis, derived from the mean values of repeated test cycles, confirms that the composite architecture serves as a viable, mechanically robust alternative to conventional metallic end shields.

1. Introduction

1.1. Industrial Context and Reliability Challenges

In the contemporary industrial landscape, electric motors driven by Variable-Frequency Drives (VFDs) represent a definitive standard, enabling significant energy savings and high-precision control of technological processes [1]. While these drives offer superior operational flexibility, extensive research on the reliability of electrical machines indicates that rolling-element bearings remain the most vulnerable element of many electromechanical drive systems. Statistical data suggests that these components account for a substantial portion of motor failures, often necessitating costly downtime and repairs [2,3].

1.2. Mechanisms of Inverter-Induced Bearing Damage

The primary cause of accelerated bearing degradation in VFD-fed systems is the introduction of high-frequency semiconductor switching. These drives generate voltage pulses with extremely steep rising edges, which produce common-mode voltages in the stator windings. These voltages are subsequently transferred to the rotor shaft through parasitic capacitances inherent within the machine structure [4,5]. As a consequence, electrical potential differences develop between the shaft and the grounded stator frame, often forcing discharge currents through the rolling bearings [6,7].

This parasitic current flow can produce electrical discharge machining (EDM) phenomena within the bearing contacts. Localized electrical discharges across the lubricant film generate microscopic arcs that melt bearing surfaces and gradually produce pitting, frosting, and characteristic fluting damage patterns [6,8,9]. These degradation mechanisms typically manifest as increased vibration levels, acoustic noise, and ultimately premature bearing failure [9,10].

Consequently the economic impact of such failures has driven the development of modern condition monitoring (CM) and predictive maintenance (PdM) strategies that aim to detect incipient faults before catastrophic damage occurs [11,12].

1.3. Classification of Mitigation Technologies

Existing mitigation techniques can be classified into several technological approaches depending on how they address the current path. One common strategy involves diverting the current away from the bearing using shaft-grounding rings or employing conductive lubrication systems to maintain a low-impedance path [4,5,13]. Alternatively, insulation-based technologies seek to interrupt the circuit entirely through the use of insulated bearings or ceramic rolling elements. While effective, these solutions often increase costs or exhibit limited long-term durability under harsh industrial conditions. In recent years, structural composite designs have emerged as a promising alternative, exploring the use of polymer-based frames or motor housings with metallic inner sleeves for bearing support [14,15]. Unlike traditional methods that focus solely on the bearing component itself, these structural approaches attempt to integrate insulation into the motor’s mechanical architecture. However, maintaining the necessary mechanical rigidity while ensuring a complete galvanic break remains a significant engineering challenge [16,17].

1.4. Proposed Segmented CFRP Architecture

Recent developments in materials science have opened new possibilities for overcoming the limitations of previous structural solutions. Fiber-reinforced composite materials are increasingly applied in mechanical structures due to their favorable strength-to-weight ratio, corrosion resistance, and design flexibility [18]. Furthermore, adhesively bonded composite joints further enable efficient load transfer while maintaining electrical insulation between structural elements [19,20]. Motivated by these developments, the present work introduces a novel segmented composite end shield. The proposed design is uniquely distinguished by its segmented 24-lamella architecture and a specific adhesive potting methodology. The final assembly of the shield is shown in Figure 1. This approach achieves a targeted galvanic break between the outer flange and the bearing housing while simultaneously fulfilling the stringent structural load-bearing requirements and electrical insulation functions within the motor assembly.

2. Materials and Methods

2.1. Theoretical Background

The design of the segmented composite end shield is predicated on two converging physical principles: the mechanical load sharing of anisotropic lamellae and the dielectric isolation provided by thin-film polymer barriers. The use of 24 discrete CFRP lamellae is designed to approximate the radial stiffness of a solid metallic disk while introducing discontinuities in the electrical path. Theoretically, the load transfer in this assembly is governed by the shear lag effect within the adhesive joints. Each lamella acts as a structural bridge where the total radial load F_r is distributed across n elements. By utilizing a balanced laminate layup (0°/90° core), the design ensures that the longitudinal modulus E_L dominates the radial stiffness, minimizing rotor eccentricity under operational loads.

From an electrical perspective, the end shield functions as a high-impedance barrier in series with the motor’s parasitic capacitances. The insulation effectiveness is defined by the volume resistivity of the epoxy matrix and the thickness of the adhesive bond line. Unlike traditional insulation, the CFRP lamellae exhibit anisotropic conductivity; while the carbon fibers are highly conductive, the epoxy encapsulation creates a dielectric barrier. The structural capacitance C_shield of the assembly is a critical parameter, as a lower capacitance further limits the high-frequency dv/dt currents that can pass through the bearing.

The reliability of the end shield depends on the fatigue life of the adhesive interfaces. Under cyclic loading, the bond line experiences fluctuating shear stresses τ. According to the theory of viscoelasticity in polymers, these joints are susceptible to micro-crack propagation at stress concentrations near the edges of the milled slots. The experimental evaluation of the S-N (stress vs. number of cycles) behavior is therefore essential to validate that the operational stresses remain below the fatigue limit of the structural epoxy.

2.2. Design and Fabrication of the Segmented End Shield

The end shield comprises 24 CFRP lamellae adhesively bonded into radial slots to connect the outer aluminum flange and the central bearing housing, providing mechanical load transfer while maintaining galvanic isolation. The initial workpiece for manufacturing the shield was aluminum alloy EN AW-6060 (AlMgSi) (NPS PROAL s.r.o., Ostrava-Svinov, Czech Republic) according to an EN 573-3 cylinder [21], from which both the outer flange with its mounting holes and the central bearing housing were machined. This machining process included milling radial slots for seating the composite lamellae.

The lamella components were fabricated from high-grade industrial carbon fiber-reinforced polymer (CFRP) panels, utilizing a specialized epoxy resin matrix. The internal architecture of these panels featured a balanced laminate layup designed for multi-directional strength. Surface Layers: A structural twill weave existed on both the top and bottom faces for esthetic finish and torsional rigidity. Core: A Unidirectional (UD) fiber core with a 50/50 distribution between 0° and 90° orientations was used to ensure optimal longitudinal and transverse stiffness. Individual lamellae—measuring 60 mm × 26 mm × 5 mm—were precision-machined from the master panels using waterjet cutting to maintain edge integrity and prevent thermal delamination [18].

It is important to note that the carbon fiber-reinforced polymer (CFRP) lamellae do not inherently provide a non-conductive barrier due to the anisotropic electrical conductivity of the carbon fibers. This is particularly critical due to the manufacturing process; the lamellae were manufactured using waterjet cutting, which shears the protective polymer matrix and leaves the bare, highly conductive carbon fibers exposed along the cut edges. Consequently, the insulating barrier within the structure relies exclusively on the adhesive layer located in the mounting grooves. To ensure strict galvanic isolation and prevent any direct contact between these exposed fibers and the metal, a two-stage encapsulation and bonding process was implemented. First, both ends of the CFRP lamellae were dipped into 3M™ Scotch-Weld™ DP490 (3M Company, St. Paul, MN, USA) structural epoxy adhesive and allowed to fully cure, creating a primary dielectric shell. Subsequently, these pre-coated lamella ends were inserted into the designated grooves of the flange and bearing housing, which were actively filled with the same DP490 adhesive. This redundant potting method guarantees a continuous, defect-free insulating barrier between the conductive CFRP and the metallic motor components [19,20,22]. After the adhesive joints were cured, the sacrificial connecting part of the workpiece was milled away. This procedure results in the final precisely centered end shield where the outer aluminum alloy flange and the central bearing housing are galvanically isolated yet rigidly joined by the carbon lamellae. Figure 2, Figure 3 and Figure 4 illustrate the key modeling stages for the subsequent fabrication and assembly of the composite end shield.

2.3. Adhesive Strength and Fatigue Testing

Structural adhesives are widely used in composite engineering applications because they enable efficient load transfer without introducing stress concentrations associated with mechanical fasteners [23,24]. To determine the optimal bonding agent, comparative adhesion tests were performed on 3M™ Scotch-Weld™ DP490 and X 60 methyl methacrylate (Hottinger Brüel & Kjær, Darmstadt, Germany). Two aluminum alloy specimens (80 × 50 × 30 mm) were utilized, each featuring bilateral grooves machined to accommodate a 5 mm wide composite lamella. Following surface degreasing, the lamellae were bonded, and upon completion of the curing process (Figure 5a,b), the specimens were subjected to static and cyclic loading protocols.

The key geometric and material parameters of the structural model are summarized in

Table 1. Geometric and material parameters of the structural model.

Parameter	Symbol	Value	Unit
Number of lamellae	n	24	-
Lamella length	L	60	mm
Lamella width	b	26	mm
Lamella thickness	t	5	mm
Adhesive layer thickness	tadh	~0.2–0.3	mm
Material (lamella)	-	CFRP	-
Material adhesive	-	DP490	-
Outer flange material	-	EN AW-6060	-

.

As a preliminary screening, a pilot series of three loading tests was performed for each adhesive variant. The results for the X 60 epoxy adhesive showed an average ultimate tensile load of 6 kN. These initial findings were compared against an identical set of specimens bonded with 3M™ Scotch-Weld™ DP490 to determine the most suitable candidate for the final design. The results of the static loading test, presented in Figure 6, clearly demonstrate significantly higher strength values and that was the reason for selecting 3M™ Scotch-Weld™ DP490 as the adhesive for the final design.

The tensile strength values for both samples are presented in the following

Table 2. Tensile force from comparison measurement.

Sample	Tensile Force [kN]
3M™ Scotch-Weld™ DP490	7.8
X 60 epoxy	6.2

and

Table 3. Tensile strength from comparison measurement.

Parameter	3M Scotch Weld DP 490	Epoxy X 60
Slot depth	5.5 mm	5.5 mm
Total bond area	286 mm2	286 mm2
Max tensile force	7.8 kN	6.2 kN
Avg. shear stress	27.27 MPa	21.67 MPa

.

The measured values for 3M Scotch Weld DP 490 correspond to typical values of the shear strength of a composite lamella glued into a groove using a structural adhesive; typical values for such systems range roughly between 10 and 35 MPa depending on the type of adhesive, surface preparation and lamella material; therefore, 27.27 MPa is within the expected range [19,20].

Based on static tests, 3M™ DP490 adhesive was selected for bonding the slats after comparative static tensile tests demonstrated its 25% higher strength than X 60 adhesive. Therefore, further cyclic fatigue tests were performed using only 3M™ DP490 adhesive.

The above-mentioned specimen consisting of two aluminum alloy box-shaped elements bonded via a composite lamella using 3M™ Scotch-Weld™ DP490 epoxy adhesive was subjected to cyclic fatigue loading. The test was conducted under sinusoidal force excitation with a constant frequency of 10 Hz. The loading amplitude was incrementally increased in three stages, as summarized in

Table 4. Cyclic loading conditions and results.

Loading level	Force Amplitude [kN]	Number of Cycles	Result
Level 1	1	2,500,000	No visible damage
Level 2	2	300,000	No visible damage
Level 3	4	250	Specimen failure

.

At Level 1, the specimen withstood 2.5 × 10⁶ cycles at 1 kN amplitude without exhibiting visible degradation or failure. Subsequently, the load amplitude was increased to 2 kN (Level 2), where the specimen endured 3 × 10⁵ cycles without detectable damage. Finally, at Level 3, the applied force amplitude was raised to 4 kN, resulting in complete bond failure after 250 cycles. Figure 7 shows the test specimen secured in the fatigue testing system.

This progressive increase in cyclic load shown in Figure 8 demonstrates the fatigue endurance of the bonded aluminum composite configuration at low to moderate amplitudes, as well as its critical threshold beyond which adhesive and/or cohesive failure occurs rapidly.

3. Results

3.1. Test of Measurement of Static Lateral Force on Rotor Shaft

The development of alternative end-shield designs for electric machines requires not only electrical validation but also thorough mechanical verification. In rotating machinery, end shields serve as critical components: they provide accurate bearing seating, ensure rotor–stator alignment, and contribute significantly to the global stiffness and dynamic behavior of the machine [24]. Any modification of this component—particularly the introduction of composite elements—may influence load transfer paths, bearing support rigidity, damping characteristics, and ultimately vibration response.

A static test was conducted to compare the stiffness of a conventional steel end shield with the newly proposed composite design. The test stand utilized two 11 kW Siemens 1LE1 series induction motors both under a nominal 400 V/50 Hz delta configuration to deliver a rated torque at 1470 rpm with an efficiency of 89.8%. One of them served as the primary drive unit. Both motors featured a robust IP55 enclosure and were equipped with 6209-2ZC3 (Brno-Hanusovice, Czech Republic) bearings at both the drive and non-drive ends. These bearings are housed within the experimental end shields, which were the subject of the structural analysis. One of these motors was equipped with standard steel end shields and the other additionally featured end shields with composite lamellae. Both units were rigidly mounted to a base.

An overall view of the motor testing and loading assembly is provided in Figure 9.

During the static test, a tensile force was applied directly to the 42 mm diameter shaft, which was measured by a force sensor at a distance of 70 mm from the motor end shield. The lateral force ranging from 250 N to 1500 N was incrementally applied to the rotor shaft. Simultaneously, the deflections of the rotor shaft and the displacement of the end shield itself were recorded.

Figure 10 shows the measured lateral compliance of the shaft-bearing–end-shield assembly; linear regression over 500 N to 1500 N was used to estimate the effective stiffness under the operating preload.

Table 5. Linear regression results (500 N–1500 N).

Parameter	Siemens Shield (Ref.)	Composite Shield
Linear Range	500–1500 N	500–1500 N
Calculated Stiffness (k)	53.12 N/µm	42.68 N/µm
Coef. of Determin. (R2)	0.998	0.942
Standard Error	0.0005	0.0021

presents the data obtained from measuring the lateral force applied to the rotor shaft.

The non-linear displacement observed in the initial loading phase of the composite end shield suggests a mechanical seating, which must be accounted for in the dynamic behavior of the shield. This settling means that while a 500 N preload applied, the composite shield is 19.6% less stiff than the Siemens metallic shield, proving that the composite’s performance is much closer to the OEM part once mechanical seating is achieved. The Siemens shield shows near-perfect linearity (R2 = 0.998). The composite shield shows slightly higher variance (R2 = 0.942), which is typical for fiber-reinforced polymers or layered composites where micro-strains and local resin compression can cause small deviations in the slope. Experimental results demonstrate that the prototype composite end shield exhibits a static stiffness of 42.68 N, which is approximately 19.6% lower than that of the standard Siemens metallic shield (57.69 N). However, under peak load conditions (1500 N), the impact on total shaft displacement is attenuated, resulting in only a 5.5% increase in total compliance compared to the OEM configuration.

3.2. Experimental Modal Analysis

To evaluate the dynamic characteristics and structural integrity of the design, experimental modal analysis (EMA) was performed using the impact-synchronous method. Structural excitation was provided via a calibrated impulse hammer, while the resulting vibration response was captured across a predefined measurement grid to facilitate high-resolution mode shape reconstruction.

The excitation grid was organized into a radial matrix consisting of 24 impact points distributed across three concentric circles (8 points per circle). To accurately capture the interaction between the isolated elements, the points were distributed as follows:

• Outer Mounting Flange: One set of eight points.
• Inner Bearing Housing: Two sets of eight points each.

As shown in Figure 11, this specialized grid configuration ensures that the relative motion between the central housing and the outer flange—joined by the carbon lamellae—is fully resolved across the frequency spectrum. The standard steel end shield weight was 4.57 kg and the composite end shield weight was 3.8 kg.

Fiber-reinforced polymer structures frequently exhibit small deviations from ideal linear elastic behavior due to internal micro-strain redistribution and local matrix compression effects [18]. The analysis revealed distinct differences in the vibration signatures of the two materials. At lower frequency regimes (725 Hz for steel and 1181 Hz for the composite), both structures exhibited complex, higher-order modes characterized by multiple nodal lines and intricate deformation patterns. However, as the excitation frequency increased, a clear modal convergence was observed; both shields transitioned toward a fundamental umbrella mode, where the geometric center experiences maximum displacement relative to the edges. Notably, the composite shield demonstrated a more uniform and distinct transition to this regime at 2193 Hz, compared to 1641 Hz for the steel reference. This significant upward shift in eigenfrequencies, alongside the variation in modal complexity, confirms that the composite’s superior stiffness-to-weight ratio—driven by its specific density and Young’s modulus—fundamentally alters the dynamic behavior and damping characteristics of the motor housing. Figure 12, Figure 13, Figure 14 and Figure 15 compare the modal shapes and frequency response of the standard metallic versus the newly developed composite shield.

The structural transition from a 4.57 kg steel end shield to a 3.80 kg composite prototype resulted in a 16.8% total mass reduction. Despite the lower absolute static stiffness of the composite material, the specific static stiffness remained comparable to the OEM reference (11.23 vs. 11.62 N/µm·kg), indicating high material utilization efficiency. Furthermore, the composite assembly exhibited a significantly lower mass participation factor (35.0% vs. 62.8%), contributing to a higher Specific Dynamic Stiffness of 19.58 N/µm·kg. This represents a 63.8% improvement in weight-adjusted dynamic performance (

Table 6. Linear regression results (500 N–1500 N).

Parameter	Steel Shield (Ref.)	Composite Shield
Physical Mass (mp)	4.57 kg	3.80 kg
Modal Mass (mm)	2.87 kg	1.33 kg
Mass Particip. (mm/mp)	62.8%	35.0%
Static Stiffness (kstat)	53.12 N/µm·kg	42.68 N/µm·kg
Spec. Static Stiffness	11.62 N/µm·kg	11.23 N/µm·kg
Spec. Dynamic Stiffness	11.95 N/µm·kg	19.58 N/µm·kg

). These results demonstrate that the composite end shield provides a superior strength-to-weight ratio under operational vibration conditions, effectively shifting critical resonances while providing enhanced damping and weight benefits for the overall test stand assembly.

3.3. Electrical Insulation Performance

In inverter-fed induction machines, common-mode voltages generated by pulse-width modulation switching can induce shaft voltages capable of driving parasitic bearing currents [5,6].

Electrical discharge phenomena associated with such currents represent a major reliability concern in modern motor drives. When supplied by a frequency converter, an 11 kW induction motor exhibits bearing voltages sufficiently low that neither electrical discharges nor measurable bearing currents occur. This is generally not the case for higher-power machines [7,9].

To test how composite shields eliminate bearing currents, the shield’s dielectric properties were first assessed. Insulation resistance and capacitance were measured using an APPA 605 insulation tester and an APPA 703 LCR meter, respectively. The objective of the measurements was to quantify the DC insulation resistance and the structural capacitance between the electrically separated aluminum flange and bearing housing, which are interconnected exclusively via carbon fiber composite lamellae and adhesive layers. The DC insulation resistance was measured using a high-voltage insulation tester APPA 605 capable of applying a controlled test voltage up to 1000 V DC. During the measurement, one terminal of the tester was connected to the outer aluminum mounting flange, while the second terminal was connected to the central bearing housing. Care was taken to ensure clean electrical contacts and to eliminate surface contamination effects.

A test voltage of 1000 V DC was applied for a duration of 60 s to allow polarization effects within the composite structure to stabilize. The leakage current was monitored internally by the instrument, and the insulation resistance was calculated according to Ohm’s law:

R = U/I

(1)

where U is the applied DC voltage and I is the measured leakage current.

The resulting insulation resistance exceeded 1 GΩ, indicating effective galvanic separation and suppression of low-frequency circulating current paths.

The structural capacitance of the composite end shield was measured using an APPA 703 precision LCR meter. The measurement was performed between the same conductive elements (flange and bearing housing). The test frequency was within the range of 100 Hz to 100 kHz with a small-signal AC excitation voltage (typically 0.5–1 V RMS) to avoid dielectric stress or non-linear polarization effects. The measured capacitance of 2.47 nF characterizes the high-frequency coupling potential between the isolated conductive parts and provides an estimate of the displacement current magnitude under inverter-induced common-mode voltage excitation. The experimental characterization of the end shield’s electrical properties was conducted using precision instrumentation from APPA Technology Corp. (Taipei, Taiwan). Structural capacitance was determined using an APPA 703 precision LCR meter, which provides a basic measurement accuracy of 0.2% and a 20,000-count resolution to ensure high-fidelity data acquisition. The galvanic isolation and insulation resistance were verified using an APPA 605 insulation tester, featuring a basic accuracy of ±(1.5% + 5 digits) for resistance measurements up to 20 GΩ. This combination of high-precision devices ensures that the reported dielectric parameters and resistance values are within strict metrological tolerances, providing a reliable basis for the subsequent analysis of current suppression.

As a final test to demonstrate the elimination of bearing currents on the operating machine, the injected voltage method test was applied.

A 12 V, 50 Hz AC source was connected with one terminal to the machine frame (ground). The second terminal was connected to the rotating shaft via a current-limiting resistor and a metallic brush providing sliding electrical contact with the rotating shaft. The shaft voltage was measured using a voltage probe, and the current was measured using a current probe. Both signals were recorded using a digital oscilloscope. See the Figure 16.

The current flowing through the bearings of a standard motor would also be sinusoidal if the brush–shaft contact were ideal and no electrical discharge phenomena occurred within the bearing. However, since neither of these conditions were met, the recorded current waveform was significantly distorted. This is the blue curve seen at Figure 17.

In the motor equipped with composite shields, the applied shaft voltage did not result in any measurable current (see dashed blue waveform) below measurement resolution (e.g., <1 µA). These results confirm that the proposed composite end-shield design effectively eliminates the conductive path responsible for parasitic bearing currents, thereby validating its electrical insulation performance.

As shown in Figure 18, the capacitive reactance X_C of the shield exhibits a characteristic decline with increasing frequency, starting from 100 MΩ near DC and dropping toward 1 kΩ at 50 kHz. Crucially, because the induced shaft voltage potential decreases at a similar rate, the total capacitive leakage current is stabilized at a negligible level (<1 µA). At higher frequencies the flange-housing impedance becomes predominantly capacitive. The measured structural capacitance of 2.47 nF suggests a theoretical path for high-frequency displacement currents; however, their operational magnitude remains negligible within the context of the drive system. In accordance with Electromagnetic Compatibility (EMC) standards, the variable-frequency drive (VFD) utilized in this study incorporates output filtration designed to attenuate switching frequencies and higher-order harmonic products. Because these filters effectively suppress the high dv/dt transients associated with frequencies above 1 kHz, the influence of parasitic capacitive currents is significantly diminished. Consequently, these results confirm that the dominant mechanism for bearing current suppression in the proposed segmented end shield is resistive galvanic isolation (>1 GΩ) rather than capacitive blocking (Figure 18). This ensures that even with a measurable residual capacitance, the high impedance of the composite–adhesive interface provides a robust barrier against circulating currents [25].

3.4. Structural Dynamic Validation of Experimental End Shields

To evaluate the dynamic characteristics and structural integrity of the newly proposed composite end shield relative to the conventional steel component, the motor equipped with experimental end shields incorporating carbon lamella elements was subjected to vibration measurements during run-up to nominal operating speed. Particular attention was paid to the vibration response during run-up, where resonance crossings and changes in excitation mechanisms may reveal potential weaknesses in stiffness, damping, or structural coupling. The measurements were performed to verify that the modified end-shield configuration does not introduce adverse vibrational characteristics compared to a conventional design (Figure 19a,b). The results therefore serve to evaluate the suitability of the composite end shields as a structural component from the perspective of vibration performance and operational reliability.

In vibration diagnostics (condition monitoring), the frequency range of interest is determined by the fundamental rotational frequency and the specific mechanical and electrical components of the motor. Figure 20, Figure 21, Figure 22 and Figure 23 depict the vibration signatures at nominal speed, comparing the frequency spectra for both composite end shields. For a four-pole induction motor operating at a nominal speed of 1470 min⁻¹, the analysis is typically divided into three primary zones [25].

Total vibration was measured across three axes: 1.26 m/s² (horizontal, x), 1.32 m/s² (vertical, y), and 2.13 m/s² (axial, z).

4. Discussion

The mechanical validation presented in this study was primarily conducted under laboratory ambient conditions, which represents a common limitation of early-stage prototype verification. Nevertheless, the structural integrity of the composite end shield was evaluated through a series of static and cyclic mechanical tests as well as extended operational measurements. In particular, long-duration run-up and steady-state operating tests were performed with the motor under mechanical load. During these experiments, the composite end shield and the bonded lamellar joints were subjected to realistic operational forces transmitted through the rotor shaft and bearings. No structural damage, degradation of adhesive joints, or loss of alignment between the bearing housing and the outer flange was observed during these tests. Furthermore, no measurable changes in vibration behavior or structural response were detected, indicating stable mechanical performance of the bonded composite structure during prolonged operation.

At the same time, it is recognized that the thermomechanical environment inside industrial electric machines can be significantly more demanding than laboratory conditions. Electric motors of the investigated power class may experience elevated temperatures during continuous operation, which can influence both adhesive mechanical properties and the stress state within hybrid structures composed of materials with different coefficients of thermal expansion. Although the selected adhesive demonstrates high mechanical strength and fatigue resistance under the tested conditions, its long-term behavior under combined thermal and cyclic mechanical loading requires further investigation. Future work will therefore focus on the evaluation of thermally induced stresses within the lamellar structure, particularly those arising from the mismatch between the thermal expansion of aluminum components and the near-zero longitudinal thermal expansion of carbon fiber composites. Planned experiments will include elevated-temperature operation and controlled thermal cycling in order to assess potential effects such as adhesive creep, changes in structural stiffness, and long-term degradation of bonded interfaces.

In parallel with these investigations, the composite end shield provides an opportunity to explore its use as an instrumented structural component for advanced condition monitoring. Modern machine diagnostics commonly relies on vibration analysis, electrical signature analysis, and advanced signal processing techniques to detect early signs of machine degradation [10,25]. However, vibration signals in inverter-fed motors frequently exhibit broadband characteristics that require complex signal processing for reliable fault identification [11]. Future research will therefore investigate the integration of a distributed network of Fiber Bragg Grating (FBG) sensors embedded or surface-mounted on the composite lamellae. Because the strain state of the lamellae is directly related to the transmitted radial and axial loads, the measured strain field could provide indirect information about rotor dynamics, bearing loading conditions, and the development of mechanical anomalies. The combination of thermomechanical validation and the development of an integrated optical sensing system represents a promising direction for future research on multifunctional composite motor components.

5. Conclusions

This study demonstrates that the proposed lamella-based composite end shield is a structurally reliable and electrically effective alternative to conventional metallic designs. By utilizing a segmented architecture of 24 CFRP lamellae, the design successfully reconciles high mechanical stiffness with superior galvanic isolation.

The primary findings of this research are as follows:

• Mechanical Integrity: Static and fatigue testing confirmed that the composite configuration maintains radial and axial rigidity comparable to that of steel components. The bonded interfaces withstood millions of cycles without degradation, while experimental modal analysis confirmed a favorable stiffness-to-weight ratio with no adverse resonances during motor operation.
• Electrical Isolation: High-voltage testing (1 kV) revealed an insulation resistance exceeding 1 GΩ and a structural capacitance of 2.47 nF. These results prove that the two-stage adhesive encapsulation effectively seals the waterjet-cut CFRP edges, creating a high-impedance barrier that suppresses parasitic circulating currents in VFD-driven systems.
• Frequency Response: Analysis of the shield’s reactance showed that even as X_C decreases at higher frequencies, the resulting capacitive current remains at a negligible level (<1 µA), providing inherent protection against modern high dv/dt switching.

In summary, the transition from a monolithic metallic shield to an engineered composite–adhesive assembly addresses critical reliability concerns without compromising vibrodiagnostic performance. This topology inherently provides a scalable platform for future “smart” integration—specifically through embedded FBG sensors—transforming a passive protective component into an intelligent node for advanced predictive maintenance.